Receipts are the stored raw model answers behind an AI visibility score — the verifiable evidence that a reported rank reflects what a model actually said.
Most AI visibility tools report a score and ask you to trust it. A receipt is the alternative: the full recorded answer — model, date, question, response text, citations — linked from the datapoint it produced. If a dashboard says your brand ranked #3 on Claude last week, the receipt is the Claude answer where that happened.
Receipts make disputes resolvable and methodology auditable. They are to AI rankings what trade records are to a market: the reason the closing number can be trusted.